topic_report: Overview visualization of multiple topics

In agoldst/dfrtopics: Tools for exploring topic models of text

Overview visualization of multiple topics

Description

This function is intended for a quick survey of a topic model. It generates a folder of image files, one for each topic. Each file has two plots, the plot of the most heavily-weighted words in the topic and the plot of the topic's relative weight in the corpus over time. The plots are saved to output_dir/<NNN>.png. If you have more than 1000 topics, other parts of this package will probably break too.

Usage

topic_report(
  m,
  output_dir = "topic_report",
  topics = 1:n_topics(m),
  breaks = "years",
  w = 1200,
  h = 800
)

Arguments

m	dfr_lda object
output_dir	directory to save plots to
breaks	time intervals by which to slice document-topic distributions for the box plots: passed on to plot_series.
w	PNG width (pixels)
h	PNG height (pixels)
topic	topics to generate plots for (by default, all)

See Also

plot_top_words, plot_series, export_browser_data and the dfr-browser web-browser-based topic model explorer

agoldst/dfrtopics documentation built on July 15, 2022, 4:13 p.m.